| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.638554216867469, | |
| "global_step": 12000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 0.0, | |
| "loss": 36.5036, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.96e-05, | |
| "loss": 8.5484, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_cer": 1.0, | |
| "eval_clean_cer": 1.0, | |
| "eval_loss": 3.6234467029571533, | |
| "eval_runtime": 384.6662, | |
| "eval_samples_per_second": 19.9, | |
| "eval_steps_per_second": 1.245, | |
| "eval_wer": 1.0, | |
| "learning_rate": 4.96e-05, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 4.792887029288703e-05, | |
| "loss": 3.2275, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_cer": 0.7080718890168877, | |
| "eval_clean_cer": 0.653960461430081, | |
| "eval_loss": 2.2960469722747803, | |
| "eval_runtime": 383.0083, | |
| "eval_samples_per_second": 19.987, | |
| "eval_steps_per_second": 1.251, | |
| "eval_wer": 0.9997886281969985, | |
| "learning_rate": 4.792887029288703e-05, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "learning_rate": 4.5841004184100424e-05, | |
| "loss": 0.9955, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_cer": 0.43273425819260275, | |
| "eval_clean_cer": 0.37557605377480335, | |
| "eval_loss": 1.2224403619766235, | |
| "eval_runtime": 382.1035, | |
| "eval_samples_per_second": 20.034, | |
| "eval_steps_per_second": 1.254, | |
| "eval_wer": 0.9549127684828382, | |
| "learning_rate": 4.5841004184100424e-05, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 4.37489539748954e-05, | |
| "loss": 0.66, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_cer": 0.36509166891345374, | |
| "eval_clean_cer": 0.304027973464835, | |
| "eval_loss": 0.9558963775634766, | |
| "eval_runtime": 380.573, | |
| "eval_samples_per_second": 20.114, | |
| "eval_steps_per_second": 1.259, | |
| "eval_wer": 0.9231907386631546, | |
| "learning_rate": 4.37489539748954e-05, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 4.1665271966527195e-05, | |
| "loss": 0.546, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "eval_cer": 0.3585320718645065, | |
| "eval_clean_cer": 0.2825745567688153, | |
| "eval_loss": 0.920676589012146, | |
| "eval_runtime": 379.317, | |
| "eval_samples_per_second": 20.181, | |
| "eval_steps_per_second": 1.263, | |
| "eval_wer": 0.948051314570021, | |
| "learning_rate": 4.1665271966527195e-05, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "learning_rate": 3.957740585774059e-05, | |
| "loss": 0.4459, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "eval_cer": 0.29398698497512193, | |
| "eval_clean_cer": 0.23832115181401903, | |
| "eval_loss": 0.7700594067573547, | |
| "eval_runtime": 382.0036, | |
| "eval_samples_per_second": 20.039, | |
| "eval_steps_per_second": 1.254, | |
| "eval_wer": 0.869307188267239, | |
| "learning_rate": 3.957740585774059e-05, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 3.748953974895398e-05, | |
| "loss": 0.4041, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "eval_cer": 0.2948846785460428, | |
| "eval_clean_cer": 0.26338499471644944, | |
| "eval_loss": 0.7755610346794128, | |
| "eval_runtime": 382.5716, | |
| "eval_samples_per_second": 20.009, | |
| "eval_steps_per_second": 1.252, | |
| "eval_wer": 0.8223663886314488, | |
| "learning_rate": 3.748953974895398e-05, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "learning_rate": 3.5397489539748955e-05, | |
| "loss": 0.3637, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "eval_cer": 0.20644807470771342, | |
| "eval_clean_cer": 0.18069009040742046, | |
| "eval_loss": 0.6014806032180786, | |
| "eval_runtime": 383.3843, | |
| "eval_samples_per_second": 19.967, | |
| "eval_steps_per_second": 1.249, | |
| "eval_wer": 0.7014942360535259, | |
| "learning_rate": 3.5397489539748955e-05, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.61, | |
| "learning_rate": 3.330962343096235e-05, | |
| "loss": 0.334, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.61, | |
| "eval_cer": 0.19067869310522317, | |
| "eval_clean_cer": 0.1638084125865915, | |
| "eval_loss": 0.5615196228027344, | |
| "eval_runtime": 378.29, | |
| "eval_samples_per_second": 20.236, | |
| "eval_steps_per_second": 1.266, | |
| "eval_wer": 0.667495894509211, | |
| "learning_rate": 3.330962343096235e-05, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "learning_rate": 3.1225941422594144e-05, | |
| "loss": 0.3283, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "eval_cer": 0.20923613323855977, | |
| "eval_clean_cer": 0.18034519196900317, | |
| "eval_loss": 0.6205064058303833, | |
| "eval_runtime": 383.8788, | |
| "eval_samples_per_second": 19.941, | |
| "eval_steps_per_second": 1.248, | |
| "eval_wer": 0.707282571581874, | |
| "learning_rate": 3.1225941422594144e-05, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 4.42, | |
| "learning_rate": 2.914644351464435e-05, | |
| "loss": 0.3762, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.42, | |
| "eval_cer": 0.1777892840510797, | |
| "eval_clean_cer": 0.1600145297640014, | |
| "eval_loss": 0.7516510486602783, | |
| "eval_runtime": 383.1576, | |
| "eval_samples_per_second": 19.979, | |
| "eval_steps_per_second": 1.25, | |
| "eval_wer": 0.63658683316261, | |
| "learning_rate": 2.914644351464435e-05, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.82, | |
| "learning_rate": 2.705439330543933e-05, | |
| "loss": 0.4954, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.82, | |
| "eval_cer": 0.2023119439201941, | |
| "eval_clean_cer": 0.1735132675824821, | |
| "eval_loss": 0.937389612197876, | |
| "eval_runtime": 379.9047, | |
| "eval_samples_per_second": 20.15, | |
| "eval_steps_per_second": 1.261, | |
| "eval_wer": 0.7072988309513357, | |
| "learning_rate": 2.705439330543933e-05, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 2.496234309623431e-05, | |
| "loss": 0.5568, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "eval_cer": 0.19821564254025834, | |
| "eval_clean_cer": 0.16663731360807796, | |
| "eval_loss": 0.8858793377876282, | |
| "eval_runtime": 379.7569, | |
| "eval_samples_per_second": 20.158, | |
| "eval_steps_per_second": 1.261, | |
| "eval_wer": 0.7027462075020731, | |
| "learning_rate": 2.496234309623431e-05, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.62, | |
| "learning_rate": 2.28744769874477e-05, | |
| "loss": 0.6756, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.62, | |
| "eval_cer": 0.19202370156131277, | |
| "eval_clean_cer": 0.16278105553598685, | |
| "eval_loss": 1.0252363681793213, | |
| "eval_runtime": 381.5363, | |
| "eval_samples_per_second": 20.064, | |
| "eval_steps_per_second": 1.255, | |
| "eval_wer": 0.6802432401671463, | |
| "learning_rate": 2.28744769874477e-05, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "learning_rate": 2.0790794979079496e-05, | |
| "loss": 0.7752, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "eval_cer": 0.2309217137674943, | |
| "eval_clean_cer": 0.19081322648820007, | |
| "eval_loss": 1.12592613697052, | |
| "eval_runtime": 381.2563, | |
| "eval_samples_per_second": 20.078, | |
| "eval_steps_per_second": 1.256, | |
| "eval_wer": 0.7656537079492057, | |
| "learning_rate": 2.0790794979079496e-05, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 6.43, | |
| "learning_rate": 1.87071129707113e-05, | |
| "loss": 0.8305, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 6.43, | |
| "eval_cer": 0.3252438785264345, | |
| "eval_clean_cer": 0.26681563343900433, | |
| "eval_loss": 1.385704517364502, | |
| "eval_runtime": 381.2294, | |
| "eval_samples_per_second": 20.08, | |
| "eval_steps_per_second": 1.256, | |
| "eval_wer": 0.9028990455750125, | |
| "learning_rate": 1.87071129707113e-05, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 6.83, | |
| "learning_rate": 1.6619246861924685e-05, | |
| "loss": 1.7385, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.83, | |
| "eval_cer": 0.9234264320203927, | |
| "eval_clean_cer": 0.911353763062111, | |
| "eval_loss": 3.2320151329040527, | |
| "eval_runtime": 377.3136, | |
| "eval_samples_per_second": 20.288, | |
| "eval_steps_per_second": 1.27, | |
| "eval_wer": 0.9998374063053834, | |
| "learning_rate": 1.6619246861924685e-05, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 7.23, | |
| "learning_rate": 1.4527196652719665e-05, | |
| "loss": 2.7839, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 7.23, | |
| "eval_cer": 0.9399648275693032, | |
| "eval_clean_cer": 0.9305836855700363, | |
| "eval_loss": 3.3238463401794434, | |
| "eval_runtime": 377.6325, | |
| "eval_samples_per_second": 20.271, | |
| "eval_steps_per_second": 1.268, | |
| "eval_wer": 0.9998861844137684, | |
| "learning_rate": 1.4527196652719665e-05, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 7.63, | |
| "learning_rate": 1.2439330543933054e-05, | |
| "loss": 2.8307, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 7.63, | |
| "eval_cer": 0.9167320277457781, | |
| "eval_clean_cer": 0.9053180403898086, | |
| "eval_loss": 3.2678146362304688, | |
| "eval_runtime": 378.715, | |
| "eval_samples_per_second": 20.213, | |
| "eval_steps_per_second": 1.265, | |
| "eval_wer": 0.9997886281969985, | |
| "learning_rate": 1.2439330543933054e-05, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "learning_rate": 1.0351464435146443e-05, | |
| "loss": 2.7672, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 8.03, | |
| "eval_cer": 0.8992438540160298, | |
| "eval_clean_cer": 0.886678848185981, | |
| "eval_loss": 3.243480682373047, | |
| "eval_runtime": 381.3935, | |
| "eval_samples_per_second": 20.071, | |
| "eval_steps_per_second": 1.256, | |
| "eval_wer": 0.9995284782856121, | |
| "learning_rate": 1.0351464435146443e-05, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 8.43, | |
| "learning_rate": 8.263598326359832e-06, | |
| "loss": 2.7426, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 8.43, | |
| "eval_cer": 0.8720403686364862, | |
| "eval_clean_cer": 0.8560562698133145, | |
| "eval_loss": 3.2396023273468018, | |
| "eval_runtime": 378.6407, | |
| "eval_samples_per_second": 20.217, | |
| "eval_steps_per_second": 1.265, | |
| "eval_wer": 0.9994634408077655, | |
| "learning_rate": 8.263598326359832e-06, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 8.84, | |
| "learning_rate": 6.175732217573222e-06, | |
| "loss": 2.7608, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.84, | |
| "eval_cer": 0.8399286747224197, | |
| "eval_clean_cer": 0.8202382000704473, | |
| "eval_loss": 3.2688863277435303, | |
| "eval_runtime": 377.8277, | |
| "eval_samples_per_second": 20.261, | |
| "eval_steps_per_second": 1.268, | |
| "eval_wer": 0.9993496252215339, | |
| "learning_rate": 6.175732217573222e-06, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 9.24, | |
| "learning_rate": 4.096234309623431e-06, | |
| "loss": 2.8195, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 9.24, | |
| "eval_cer": 0.8083991029191891, | |
| "eval_clean_cer": 0.7865445579429376, | |
| "eval_loss": 3.3283286094665527, | |
| "eval_runtime": 385.569, | |
| "eval_samples_per_second": 19.854, | |
| "eval_steps_per_second": 1.242, | |
| "eval_wer": 0.9989106222460693, | |
| "learning_rate": 4.096234309623431e-06, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 9.64, | |
| "learning_rate": 2.00418410041841e-06, | |
| "loss": 2.9044, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 9.64, | |
| "eval_cer": 0.7861252726782519, | |
| "eval_clean_cer": 0.7617338851708348, | |
| "eval_loss": 3.4099173545837402, | |
| "eval_runtime": 381.2676, | |
| "eval_samples_per_second": 20.078, | |
| "eval_steps_per_second": 1.256, | |
| "eval_wer": 0.9988130660292994, | |
| "learning_rate": 2.00418410041841e-06, | |
| "step": 12000 | |
| } | |
| ], | |
| "max_steps": 12450, | |
| "num_train_epochs": 10, | |
| "total_flos": 1.2874965830163638e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |